Replacing RCTs with real world data for regulatory decision making: a self-fulfilling prophecy?

Abstract

Real world data are advocated as an alternative approach to RCTs for closing knowledge gaps on drugs, but Beate Wieseler and colleagues argue that this approach is the wrong remedy for current challenges in drug development

Very‐high frequency plasmapheresis and donor health–absence of evidence is not equal to evidence of absence

“…While RCTs are costly and time‐consuming, real‐world data via prospective cohort studies are essential for investigating the epidemiology of the health status and potential safety issues and providing as basis for regulatory and health policy decision making. However, caution is needed here as such use of observational studies from routine practice data sources might hinder efforts needed to undertake prospective randomized controlled trials and may produce results discordant with RCTs 6,7 . We advocate that RCTs together with analyses of real‐world data (via improved epidemiological designs and methods to control for biases) are complementary and serve as the most robust study designs to investigate the efficacy, effectiveness, and safety of repeated plasmapheresis 8 .…”

RCTs and real world evidence are complementary, not alternatives

“…Wieseler and colleagues discuss the use of randomised controlled trials (RCTs) and real world data in regulatory decision making 1. This is an important topic, but the article presents a distorted view of real world evidence, which is being used as a complement, not an alternative, to RCTs 2…”

Authors’ reply to Morales and Arlett

“…We read Morales and Arlett’s response1 to our article on replacing randomised controlled trials (RCTs) with real world data2 with interest. We appreciate that the European Medicines Agency now clarifies that real world data can also be collected in clinical trials, and we look forward to the updated definitions of both real world data and real world evidence.…”

“…We agree with the EMA that research questions about, for example, disease epidemiology or drug use can be answered with non-interventional real world data. But these questions were not the subject of our article 2. We discussed the investigation of treatment effects of drugs and the specific shortcomings of non-randomised, non-interventional real world data in this context.…”